{"id":1631,"date":"2023-12-21T10:04:19","date_gmt":"2023-12-21T09:04:19","guid":{"rendered":"https:\/\/babel.isa.uma.es\/kipr\/?p=1631"},"modified":"2023-12-21T10:04:42","modified_gmt":"2023-12-21T09:04:42","slug":"they-had-to-do-it-certified-rl-through-online-reward-shaping-definition","status":"publish","type":"post","link":"https:\/\/babel.isa.uma.es\/kipr\/?p=1631","title":{"rendered":"They had to do it: Certified RL (through online reward shaping\/definition)"},"content":{"rendered":"<h4>Hosein Hasanbeig, Daniel Kroening, Alessandro Abate, <strong>Certified reinforcement learning with logic guidance, <\/strong>Artificial Intelligence, Volume 322, 2023 <a href=\"https:\/\/doi.org\/10.1016\/j.artint.2023.103949\" target=\"_blank\">DOI: 10.1016\/j.artint.2023.103949<\/a>.<\/h4>\n<blockquote><p>Reinforcement Learning (RL) is a widely employed machine learning architecture that has been applied to a variety of control problems. However, applications in safety-critical domains require a systematic and formal approach to specifying requirements as tasks or goals. We propose a model-free RL algorithm that enables the use of Linear Temporal Logic (LTL) to formulate a goal for unknown continuous-state\/action Markov Decision Processes (MDPs). The given LTL property is translated into a Limit-Deterministic Generalised B\ufffdchi Automaton (LDGBA), which is then used to shape a synchronous reward function on-the-fly. Under certain assumptions, the algorithm is guaranteed to synthesise a control policy whose traces satisfy the LTL specification with maximal probability.<\/p><\/blockquote>\n","protected":false},"excerpt":{"rendered":"<p>Hosein Hasanbeig, Daniel Kroening, Alessandro Abate, Certified reinforcement learning with logic guidance, Artificial Intelligence, Volume 322, 2023 DOI: 10.1016\/j.artint.2023.103949. Reinforcement <span class=\"ellipsis\">&hellip;<\/span> <span class=\"more-link-wrap\"><a href=\"https:\/\/babel.isa.uma.es\/kipr\/?p=1631\" class=\"more-link\"><span>Read More &rarr;<\/span><\/a><\/span><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[84],"tags":[526,493,527,520],"class_list":["post-1631","post","type-post","status-publish","format-standard","hentry","category-reinforcement-learning-in-ai","tag-certified-rl","tag-model-free-reinforcement-learning","tag-reward-shaping","tag-safe-rl"],"_links":{"self":[{"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=\/wp\/v2\/posts\/1631"}],"collection":[{"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1631"}],"version-history":[{"count":1,"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=\/wp\/v2\/posts\/1631\/revisions"}],"predecessor-version":[{"id":1632,"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=\/wp\/v2\/posts\/1631\/revisions\/1632"}],"wp:attachment":[{"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1631"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1631"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/babel.isa.uma.es\/kipr\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1631"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}